THE JOURNAL OF CHEMICAL PHYSICS 130, 034705 共2009兲

Molecular dynamics simulation of the contact angle of liquids

on solid surfaces
Bo Shi and Vijay K. Dhira兲
Department of Mechanical and Aerospace Engineering, Henry Samueli School of Engineering and Applied
Science, University of California, Los Angeles, California 90095, USA
共Received 5 August 2008; accepted 3 December 2008; published online 21 January 2009兲

In this work, Lennard-Jones liquid and water droplets are simulated adjacent to a solid surface using
molecular dynamics. The contact angle is obtained by using direct simulation. The particle-particle
particle-mesh method 关B. Shi, S. Sinha, and V. K. Dhir, J. Phys. Chem. 124, 204715 共2006兲兴 for
long range force correction is used in the simulation. The contact angle is studied as a function of
system temperature and the solid/fluid interaction potential. It is shown that the contact angle
decreases with increasing system temperature and increases when the potential decreases. At high
system temperature 共pressure兲, the contact angle drops to zero. The predictions are compared with
data from experiments and a reasonable agreement is found. © 2009 American Institute of Physics.
关DOI: 10.1063/1.3055600兴

I. INTRODUCTION In this paper, the particle-particle particle-mesh 共PPPM兲

Solid-liquid-vapor interaction phenomena have an im-
portant role in phase change heat transfer. In 2005, Sinha1
carried out simulation of contact angle using argon and a
virtual solid wall. He modeled the solid-liquid interaction by
a 共9,3兲 potential and assumed that the solid surface is semi-
infinite,
method6 is applied to minimize the error in long range terms
in both the Coulombic and Lennard-Jones potentials. Higher
accuracy simulation is performed to obtain the contact angle
variation with temperature and wall potential.
II. SIMULATION MODEL

2冑2␲⑀sf ␴3sf
冋 冉 冊 冉 冊册
In the simulation, a liquid droplet is placed on a solid
␴sf 9
␴sf 3
surface. The contact angle can be obtained directly from the
⌽共z兲 = 2 − 15 , 共1兲
45R30 z z density contour of the droplet after the system reaches equi-
librium.
where R0 is the lattice constant of a fcc共111兲 solid surface, z In carrying out the simulation of the contact angle of
is the distance from wall, and ⑀sf and ␴sf are Lennard-Jones argon droplet on a virtual solid wall, the size of the simula-
parameters, respectively. In that study, it was found that the
contact angle decreased with increasing temperature and no 300
matter how the potential was changed, the trend was the
same. Earlier in 2002, Maruyama et al.2 also simulated the
contact angle of argon liquid on a virtual solid wall. They 250
calculated various contact angles by changing the parameter
⑀ in the potential between argon fluid and the virtual solid
wall. They found that the contact angle decreased with an 200
increase in interface potential and became 0° when the po-
tential was set to a large enough value.
Results of limited molecular dynamics simulation stud- 150
ies have been reported in literature on the contact angle of
water and a realistic solid wall. In 2003, Kimura and
Maruyama3 first simulated a water droplet on a platinum 100
B
surface. They used two different potentials between water
and platinum. By using Spohr–Heinzinger4 potential, they
found that the contact angle is always 0°, which suggests that
50
this potential is not suitable to predict the contact angle. By A

using Zhu–Philpott5 potential, they successfully calculated

the contact angle variation with the orientation of the crystal
structure of a platinum wall at a temperature of 350 K. 20 40 60 80 100 120 140 160 180 200

FIG. 1. 共Color online兲 Liquid droplet and the contact angle 共each unit is
a兲
Electronic mail: vdhir@seas.ucla.edu. 0.2352 Å兲.

0021-9606/2009/130共3兲/034705/5/$25.00 130, 034705-1 © 2009 American Institute of Physics

Downloaded 16 Nov 2009 to 137.54.15.70. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/jcp/copyright.jsp
034705-2 B. Shi and V. K. Dhir J. Chem. Phys. 130, 034705 共2009兲

300 300

250 250

200 200

Z 150 Z 150

100 100

50 50

20 40 60 80 100 120 140 160 180 200 20 40 60 80 100 120 140 160 180 200

(a) X (c) X

300

250

200

Z 150

100

50

20 40 60 80 100 120 140 160 180 200

(b) X

FIG. 2. 共Color online兲 Droplet contour at ␴r 共a兲 ⑀r = 0.408, Tⴱ = 0.7 共b兲 ⑀r = 0.408 Tⴱ = 1.0, and 共c兲 ⑀r = 0.65Tⴱ = 1.0 共argon兲 共each unit is 0.2352 Å兲.

tion box is set as 47.04⫻ 47.04⫻ 70.56 Å3. The solid wall is
assumed to consist of four layers of atoms arranged as a fcc
lattice with its 具111典 surface in contact with the fluid. The
solid wall is taken to be that of platinum with a lattice con-
stant of 2.776 Å and density of 21.45⫻ 103 kg/ m3. For the
solid-solid interaction, the Lennard-Jones potential is applied
with the length parameter ⑀ = ⑀ss = 8.35⫻ 10−20 J and ␴ = ␴ss
= 2.475 Å to model the platinum atomic vibrations in the
lattice. The method of tethering the solid atoms to a fixed
lattice site is not used.
The Lennard-Jones equation is also used for the solid-
liquid interaction. The potential parameters used are ␴sf
= 0.8␴ f f and ⑀sf = 0.6⑀ f f , where ␴ f f = 3.4 Å and ⑀ f f = 1.65324
⫻ 10−21 J for argon. To study the contact angle variation
with the strength of wall potential, a relative parameter ⑀r
= ⑀sf / ⑀ f f has been used.
Initially, 980 atoms are stacked in a simple cubic struc-
ture above the center plane. The initial velocities are as-
signed to the atoms randomly according to the system tem-
perature. Periodic boundary conditions are applied along the
x- and y-directions. In the z-direction, the mirror boundary
condition is used, that means, if an atom reaches the bound-
ary, the x , y velocities will remain the same while its
z-direction velocity changes sign.
The Lennard-Jones potential is used for liquid-liquid in-
teractions with ⑀ f = 120⫻ 1.3806⫻ 10−16 erg/ atom and ␴ f
= 3.4 Å. The PPPM method is used to minimize the long
range error.
In order to simulate the contact angle of a water droplet
on a platinum surface, the domain size is chosen to be 80
⫻ 80⫻ 80 nm3. Equivalent potential field is applied instead

Downloaded 16 Nov 2009 to 137.54.15.70. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/jcp/copyright.jsp
034705-3 Contact angle simulation J. Chem. Phys. 130, 034705 共2009兲

of the five layers of platinum atoms 关fcc共111兲兴 at the bottom

of simulation domain. Initially, 13⫻ 13⫻ 10 water molecules
were placed on the platinum surface.
Periodic boundary conditions were specified at the four
side surfaces and a mirror boundary at the top surface. In this
simulation, the simple point charge/extended potential7 and
Z-P potential5 were used for water-water and water-platinum
interactions, respectively. The time step was 1 fs, which is
considered to be small enough for the simulation.
At a desired temperature, Berendsen’s thermostat was
applied at every 5 time steps for the first 100 000 time steps.
The next 100 000 time steps were used for relaxation to
make sure the system reached equilibrium. After that, sam-
pling was performed every 20 time steps for the next
100 000 time steps. To eliminate the momentum introduced
by the computational error during the simulations, the center
of mass of the system was adjusted in the x-y plane by mov-
ing the simulation box to the new center of mass of the FIG. 3. Variation of the contact angle with ⑀r 共argon兲.
system after every 500 time steps without altering the rela-
tive positions of the atoms.
formed by Sinha.1 However, he applied an equivalent poten-
In the simulation, density was sampled by dividing the
tial model instead of the real solid wall. This behavior is also
simulation domain into 200⫻ 200⫻ 300 small cubes. Each
observed in experiments.
cube length represents 0.2352 Å. To determine the contact
If one fixes the solid-fluid combination 共⑀r兲, and gradu-
angle, densities were postprocessed in two center planes: x-z
plane and y-z plane. A contour line corresponding to the ally increases the temperature, the contact angle calculated
average of liquid and vapor densities was plotted, and it de- from direct simulation decreases, as is shown in Fig. 4. The
fined the boundary of the liquid droplet, as shown in Fig. 1. contact angle decreases slowly in the temperature range 0.7–
The droplet boundary shows statistical fluctuations and has a 0.9. For T ⱖ 0.9, the contact angle decreases very rapidly and
curved interface. The contact angle was calculated by draw- the solid becomes fully wetted at T = 1.0– 1.1, which is lower
ing a tangent line 共A-B兲 near the contact line such that line than the critical temperature 共Tc = 1.26兲. The interfacial ten-
A-B overlaps the contour in the 0 ⱕ z ⱕ 3␴ region. For each sion between liquid and vapor ␥lv decreases with tempera-
test case, the contact angle was measured in the x-z and y-z ture, as shown by Shi.8 From Young’s equation
planes and was averaged.
␥sv − ␥sl
cos ␪ = , 共2兲
␥ lv

III. RESULTS AND DISCUSSION it can be inferred that ␥lv approaches the difference between
A. A Lennard-Jones droplet on a solid surface the solid-vapor interfacial tension ␥sv and the solid-liquid
interfacial tension ␥sl, and a wetting condition is attained.
Molecular dynamics simulations were carried out to The contact angle of the high energy surface 共high ⑀r兲 has
study the contact angle of a droplet of Lennard-Jones fluid
on a solid wall. The dependence of contact angle on the
solid-liquid potential and the system temperature was deter-
mined both by direct simulation and by applying Young’s
equation at the nanolevel. In the simulation of argon, all the
parameters are nondimensionalized using the Lennard-Jones
parameters ␴, ⑀, and mass of argon atom, m0.
Figures 2共a兲–2共c兲 show density contours of droplets ob-
tained from direct simulation for different wall potentials and
temperatures. It is seen from Figs. 2共a兲 and 2共b兲 that the
contact angle decreases as temperature is increased. Simi-
larly, it is seen from Figs. 2共b兲 and 2共c兲 that the contact angle
decreases as ⑀r is increased from 0.408 to 0.650.
More detailed results of the variation of contact angle
with solid-liquid potential as obtained from direct simulation
are presented in Fig. 3. Here, the contact angles were evalu-
ated in both x-z and y-z planes from the droplet contour. The
difference between the contact angles determined in the two
midplanes was less than ⫾3°, and they were averaged to
obtain the mean value. Similar calculation were also per- FIG. 4. Variation of the contact angle with temperature 共argon兲.

Downloaded 16 Nov 2009 to 137.54.15.70. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/jcp/copyright.jsp
034705-4 B. Shi and V. K. Dhir J. Chem. Phys. 130, 034705 共2009兲

FIG. 5. Density distribution of argon fluid layer placed on a solid wall.

lower values at all temperatures and also approaches the wet-
ting condition at lower temperatures.
In Fig. 5, the density of argon shows fluctuations near
the solid wall. These fluctuations decay with increasing z,
and the density approaches a constant value corresponding to
the bulk liquid density. It should be mentioned that the film
thickness plays an important role in the calculation of density
and surface tension. Sinha1 suggested that the thickness for
simulation layer must be larger than 2 nm, otherwise both
density and surface tension values will differ from experi-
mental values. It can also been seen from Fig. 5 that if the
thickness is less than 2 nm, the fluctuation will still exist and
there could be large errors in the calculated values.
Near the solid wall, the density distribution is as if the
liquid is in a solid state. Small gaps can be seen between the
“quasisolid” fluid atoms. The layer that touches the solid
wall shows a very high density in comparison to the bulk
argon fluid. Such high density regions result from the dense
packing of molecules. Increased density can also lead to
higher local thermal conductivity as shown by Shukla and
Dhir.9
FIG. 6. 共Color online兲 Dependence of water-platinum contact angle on
temperature.
angle decreases with the increase in the system pressure and
temperature. This variation can be expressed as
␪ = A共Tⴱ兲, 共3兲
where A is a function of dimensionless temperature Tⴱ. The
variation of A is determined by fitting a second order poly-
nomial as
A = − 31.61 + 245.33Tⴱ − 265.44Tⴱ2 共Tⴱ ⬍ 0.75兲. 共4兲
If the temperature is high enough, the contact angle goes to
zero and the water droplet degenerates to a water film.
To study the change in contact angle with the wall po-
tential, a multiplier ␣ was used in the potential between the
solid and the liquid. The corresponding results are shown in
Fig. 7. It again shows that the contact angle increases with a
decrease in potential. As the solid potential is decreased, ␥lv
remains constant because the potential between liquid and
vapor does not change. At low pressures 共temperatures兲,

B. A water droplet on a platinum surface

Similar calculations were also performed for a water
droplet and a slab. The PPPM method was applied to mini-
mize the error in long range terms.
At first, Z-P potential is applied in the simulation. After
100 000 time steps 共100 ps兲 at a temperature of 310 K, it is
found that there is a monolayer of water all along the plati-
num surface. This is because the Z-P potential is sharp near
the platinum surface, and so the water molecules are at-
tracted to the surface. To minimize the affect of this mono-
layer, the contact angle is measured from the top surface of
the monolayer. From the density contour of the droplet, the
contact angle for water-platinum can be determined.
Figure 6 shows the dependence of the water-platinum
contact angle on the dimensionless system temperature Tⴱ
= T / Tc; Tc is the critical temperature of water 共647 K兲. In the
figure the squares are the simulation results from this study
and the circle is the result of Kimura and Maruyama3 for FIG. 7. Variation of contact angles with temperature and surface energy
one, temperature. From Fig. 6 it can be seen that the contact 共water case兲.

Downloaded 16 Nov 2009 to 137.54.15.70. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/jcp/copyright.jsp
034705-5 Contact angle simulation
FIG. 8. 共Color online兲 Comparison of water-platinum 共Z-P兲, argon-virtual
solid wall, and water-aluminum results.
change in ␥sv with temperature is smaller than that in ␥sl
since the vapor density is much smaller than the liquid den-
sity.
In Fig. 7, the variation of the contact angle of water with
temperatures is plotted for several wall potentials. The solid
symbols are results from direct simulation of droplet and the
open symbols are from Young’s equation by using the values
of ␥sl, ␥sv, and ␥lv determined from simulation of a liquid
slab on a solid surface. It is seen that for different wall po-
tentials and system temperatures, these two sets of results
again match reasonable well.
In 1981, Sullivan10 used the Van der Walls model to
study the contact angle dependence on temperature. He con-
cluded that the contact angle depends on T / Tc and showed a
trend similar to that found here. Lay and Dhir11 carried out a
theoretical study of the variation of the contact angle with
the saturation pressure and temperature for water and found
that it decreases rapidly at temperatures near critical. To
verify the observed behavior for water platinum, the results
for argon-virtual solid wall and the data of water-aluminum
obtained by Berbardin et al.12 are plotted in Fig. 8. Here, the
contact angles are nondimensionalized with ␪0, which is its
Downloaded 16 Nov 2009 to 137.54.15.70. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/jcp/copyright.jsp
J. Chem. Phys. 130, 034705 共2009兲
contact angle at the triple point 共simulated value at 273.15
K兲. It is found that water platinum shows a dependence on
temperature that is similar to argon-virtual solid wall and
water aluminum. At low temperatures, the contact angle de-
creases slowly with temperature, while at high temperatures,
the contact angle decreases rapidly to zero.
IV. CONCLUSIONS
In this work, molecular dynamics simulation was used to
study the contact angle of Lennard-Jones liquid and water on
solid surface. Using direct simulation of a liquid droplet on a
solid surface, the contact angles are determined from the
density contour of the droplet.
Variation of the contact angle with system temperature
and the solid/fluid interaction energy was also studied. It is
found that the contact angle decreases with increasing sys-
tem temperature or solid-fluid potential. When the tempera-
ture or the solid-fluid potential is high enough, the contact
angle decreases to zero. The calculated contact angles for
water platinum, argon-virtual solid wall, and the experimen-
tal value of water-aluminum cases were compared and found
to show similar trend.
ACKNOWLEDGMENT
The authors gratefully acknowledge the support this re-
search received from NASA under the Microgravity Fluid
Physics Program.
1
S. Sinha, Ph.D. dissertation, University of California, 2005.
2
S. Maruyama, T. Kimura, and M.-C. Lu, Therm. Sci. Eng. 10, 23 共2002兲.
3
T. Kimura and S. Maruyama, The Sixth ASME-JSME Thermal Engineer-
ing Joint Conference, 2003 共unpublished兲.
4
K. Heinzinger and E. Spohr, Electrochim. Acta 34, 1849 共1989兲.
5
S. B. Zhu and M. R. Philpott, J. Chem. Phys. 100, 6961 共1994兲.
6
B. Shi, S. Sinha, and V. K. Dhir, J. Phys. Chem. 124, 204715 共2006兲.
7
H. J. C. Berendsen, J. R. Grigera, and T. P. Straatsma, J. Phys. Chem. 91,
6269 共1987兲.
8
B. Shi, Ph.D. dissertation, University of California, 2006.
9
R. Shukla and V. K. Dhir, Proceedings of 2005 ASME Summer Heat
Transfer Conference, 2005 共unpublished兲.
10
D. E. Sullivan, J. Chem. Phys. 74, 2604 共1981兲.
11
J. H. Lay and V. K. Dhir, Conference Proceedings of ASME Heat Trans-
fer Division, 1995 共unpublished兲.
12
J. D. Bernardin, I. Mudawar, C. B. Walsh, and E. I. Franess, Int. J. Heat
Mass Transfer 40, 1017 共1997兲.